cultus 4 · the journal of intercultural mediation and communication corpus linguistics and culture...

________________________________________________

This is a contribution from Cultus: the Intercultural Journal of Mediation and Communication 2011: 4 © Iconesoft Edizioni Gruppo Eurosan This electronic file may not be altered in any way. The author(s) of this article is /are permitted to use this PDF file to generate printed copies to be used by way of offprints, for their personal use only.

__________________________________________________

THE JOURNAL OF INTERCULTURAL MEDIATION AND COMMUNICATION

CORPUS LINGUISTICS AND CULTURE

2011, Volume 4

Iconesoft Edizioni Terni - Italy

Cultus

CULTUS __________________________________________________

2

Registrazione al Tribunale di Terni n. 11 del 24.09.2007

Direttore Responsabile Agostino Quero

Editore Iconesoft Edizioni Anno 2011

ISSN 2035-3111

© Iconesoft Edizioni – Gruppo Eurosan Italia srl via Garibaldi 89 – 05100 Terni

La realizzazione di questo volume è stata resa possibile grazie al contributo dei Monti dei Paschi di Siena in collaborazione con

l’Università del Salento

________________________________________________________

3

CULTUS

the Journal of Intercultural Mediation and Communication

2011, Volume 4

CORPUS LINGUISTICS AND CULTURE

Editors

David Katan University of Salento and University of Trieste

Elena Manca

University of Salento

Cinzia Spinzi University of Bologna

ICONESOFT EDIZIONI TERNI

CULTUS __________________________________________________

4

CULTUS

the Journal of Intercultural Mediation and Communication

Editorial Board

Michael Agar Ethknoworks LLC and University of Maryland, College Park, USA Patrick Boylan SIETAR Milton Bennet Intercultural Development Research Institute, Italy Patrick Boylan SIETAR-Italy and past Professor at Roma Tre University, Rome Ida Castiglioni University of Milan (Bicocca), Intercultural Development Research Institute Andrew Chesterman University of Helsinki, Finland Delia Chiaro University of Bologna (SSLMIT), Forlì, Italy Nigel Ewington WorldWork Ltd, Cambridge, England Peter Franklin HTWG Konstanz University of Applied Sciences, dialogin-The Delta Intercultural Academy

________________________________________________________

5

Maria Grazia Guido University of Salento, Italy Raffaela Merlini University of Macerata, Italy Robert O’Dowd University of León, Spain. Anthony Pym Intercultural Studies Group, Universidad Rovira I Virgili, Tarragona, Spain Helen Spencer-Oatey University of Warwick, England Federica Scarpa SSLMIT University of Trieste, Italy Helen Spencer-Oatey Universoty of Warwick, England Christopher Taylor University of Trieste, Italy Kumiko Torikai Rikkyo Graduate School of Intercultural Communication, Tokyo, Japan David Trickey TCO s.r.l., International Diversity Management, Bologna, Italy Margherita Ulrych University of Milan, Cattolica del Sacro Cuore, Italy

CULTUS __________________________________________________

6

________________________________________________________

7

Table of Contents Corpus Linguistics and Intercultural communicative approach: a synergy. Introduction by Cinzia Spinzi 9

A conversation between Mike Stubbs and Sebastian Hoffmann 21 Corpus Linguistics: what it is and what it can do Alan Partington 35 What A TripAdvisor Corpus Can Tell Us About Culture Maria Elisa Fina 59 Perceptions of Europe in British and Italian TV news programmes

Silvia de Candia and Marco Venuti 81 Intercultural space in translation education: Parallel corpus as a tool for exploration

Xiaoping Jiang and Josta van Rij-Heyligers 99 Notes on contributors 120

Guidelines for contributors 123

Jiang-van Rij-Heyligers ________________________________________________________

99

Intercultural space in translation education: Parallel corpus as a tool for exploration

Xiaoping Jiang and Josta van Rij-Heyligers

Abstract

This article aims to explore the uses of parallel corpora in translation education (Chinese-English) from an intercultural perspective. In this article, a brief discussion of corpus-based methods in translation studies is provided first, followed by examples of Chinese source texts and their translated English texts to accentuate the impact of culture on the outcome of translation. To address the challenges of cultural dissimilarities in translation, an intercultural approach is advocated to emphasize the importance of the intercultural space that can be navigated and negotiated in the translation process as the target texts contain elements of both languages and cultures. Finally, the article discusses the role of parallel corpora in translation education, and identifies the current limitations as well as promising developments in corpus design for translation studies.

1. Introduction Since the 1990s, corpus-based approaches to studying translation

have received growing interest, as a wealth of data has emerged from studies using multilingual corpora, parallel corpora, and comparable corpora. Examples are the English-Norwegian and Europarl parallel

CULTUS __________________________________________________

100

corpus, and multilingual publications (comparable corpora) by international organisations (UN, WHO etc.) and news agencies (Granger, 2003; Zanettin, 1998). These corpora facilitate descriptive analyses of (multiple) translated texts and source texts, and provide valuable sources of information for translation studies and education. Since translators often spend a substantial amount of time on consulting reference materials, the use of computer-based bilingual parallel corpora can speed up the translation process, enable more native-like interpretations and strategies in both source and target texts (Aston 1999), and help trainee translators notice the general patterns and preferred ways of expressing things. In the use of translational corpora it is important, however, not to neglect the social and cultural contexts in which translations are produced: trainee translators need to be linguistically competent and aware of the specific context in which the translation act occurs. In particular, corpora that contain multiple translations of the same source text can help students gain better understanding of how translational acts are influenced by the social and cultural as well as historical and ideological context (cf. Schäffner 2004).

This paper first explores the potential use of corpus-based methods in translation studies. Then the need for an intercultural perspective is illustrated by discussing differences between Chinese and English texts at the semantic, syntactic and discourse levels. This is followed by presenting an intercultural approach to translation education so as to advance trainee translators’ skills in intercultural mediation. Finally, the use of parallel corpora for teaching translation and their current limitations as well as future potential are explored.

2. Use of corpora in translation studies The use of corpora in the field of translation originates from corpus

linguistics. The early focus was, however, on developing English monolingual corpora for lexicographical use: to analyze and describe the syntactic, semantic and paradigmatic relationships within the lexicon of a particular language; and to compile and edit English dictionaries. Corpus application in translation studies has been a more recent enterprise. With more data collected from corpus research and new developments in bilingual and multilingual corpora, corpora are becoming important


101

resources for translation studies. Mona Baker was among the first to introduce corpus-based approaches to translation studies in the early 1990s, when she compiled a corpus of translated texts to uncover “[their] nature … as a mediated communicative event” (Baker 1993: 243). Later corpus-based studies shifted the research focus from source text investigations that often led to prescriptive translation practices to a target text orientation that seeks to reveal distinct patterns of translation (Granger 2003). Baker’s work, in particular, revealed several possible translation universals that later corpus studies helped to both validate and challenge (cf. Puurtinen 2003). These universals highlight the typical features of the translated text rather than the original text, and are represented by simplification, explicitation and normalization (cf. Baker 1993).

Bilingual comparable and parallel corpora have in particular proved valuable for translation theory and practice: they provide “new insights into the languages compared - insights that are not likely to be noticed in studies of monolingual corpora”, and “increase our knowledge of language-specific, typological and cultural differences, as well as of universal features” (Aijmer and Altenberg 1996: 12). Comparable corpora contain authentic source language texts in more than one language that share a common theme or subject matter. The sub-corpora are not direct translations of each other, but are collected on the basis of the same sampling frame, and comparable balanced content and representativeness (Gupta 2008). In a parallel corpus, however, the authentic source text is matched line by line with its translated text. Such a corpus can have uni-directional source texts, as in the case of Baker’s corpus, as well as bi-directional or multi-directional source texts. For instance, English as source text only with Chinese as translated text is uni-directional, and English and Chinese as source texts as well as respective translated texts represent bi-directional corpora. The different levels of complexity can make a parallel corpus hard to compile compared to a comparable corpus (Gupta 2008). However, a comparable corpus may not serve translation studies well on its own as there is no direct comparison possible between source text and translated text. As a parallel corpus “illuminate[s] differences between source texts and translations, and between native and non-native texts” (Aijmer and Altenberg 1996: 12), it can raise awareness of the “particular translation practices and procedures … used by the translator” (Shuttleworth and Cowie 1997: 120) and/or translators in the case of multiple translations

CULTUS __________________________________________________

102

of the source text. Moreover, it can give some indication of the cross-cultural production and strategies involved in making culturally loaded language, such as expressed by collocations, idioms and metaphors, comprehensible across languages (cf. Monti 2009).

In this paper, the potential merit of parallel corpora in translation education is explored. Three Chinese source texts and phrases and their translation in English are selected as exemplars to highlight the linguistic and cultural differences embedded in the two languages. They could further be used to illustrate the challenge in constructing a parallel corpus: aligning texts where the source and target language represent very different linguistic and cultural communities, thus requiring ‘free’ adaptions of the source text for intelligibility.

3. Elements of language and culture in Chinese-English translation

Translation as text production entails that cultural elements may

hinder effective renditions of the original text. Cultural elements are often reflected at the semantic and syntactic level. They can affect successful translation and highlight the value of an intercultural approach in translation studies and education.

At the semantic level, a word can have denotative and connotative meanings. In intercultural situations, communicative problems often arise from connotative meanings since interlocutors tend to behave as if their reactions to a word or concept are the same as those somebody else would have. These problems can lead to misunderstandings and even communication breakdowns.

Take the word 红 (pronounced “hong”) in Chinese and its English counterpart ‘red’ as an example. They share the same denotative meaning, but the connotative meaning can evoke very different reactions. 红 (hong) in Chinese has more positive connotations, such as “fortune”, “success”, “wealth”, “happiness” and “popularity”, whilst ‘red’ has relatively more negative connotations in English such as “danger”, “violence”, and “disaster”. The different connotations generate different word associations and meanings, making it hard to translate the resulting word combinations. For example, 红袖 (Hong-Xiu) is a synonym of a well-educated beauty in classical Chinese


103

literature. But the word for word translation ‘red sleeve’ of Hong-Xiu loses the true meaning of the expression.

Syntactically, Chinese and English are distinct languages. Chinese has a paratactic (coordinate) sentence structure: sentences or clauses are constructed by means of semantic ties. English, on the other hand, has a hypotactic (subordinate) sentence structure: sentences or clauses are put together through linguistic forms, including conjunctions, adverbs, and prepositional phrases to create meaning. In Chinese, if meaning is understandable, the sentence is acceptable, whilst in English the focus is on form; without the proper forms, meanings cannot be clearly expressed (Song 2004). To illustrate these syntactic differences, consider the following poem by Ma Zhiyuan (马致远), a famous classical Chinese poet, and its published (free) English translation by David Hawks (Jia 1997):

Original Chinese poem

Literal Translation Free Translation

古藤，老树，昏鸦, Dry vines, old trees, evening crows,

Crows hovering over rugged old trees wreathed with, rotten vine – the day is about done.

小桥，流水，人

家， Little bridge, murmuring brook, rural cottage,

Yonder is a tiny bridge over a sparkling stream, and on the far bank, a pretty little village.

古道，西风，痩

马。 Ancient road, west wind, thin horse,

But the traveller has to go on down this ancient road, the West wind moaning, his bony horse groaning,

夕阳西下，断肠人 Sunset, broken heart, at Trudging towards the

CULTUS __________________________________________________

104

在天涯。 the end of the world. sinking sun, farther and farther away from home. (Hawks)

The first three lines in the Chinese poem comprise coordinated noun

phrases, and the last line contains no subject. Not restricted by form, the Chinese language offers great freedom of creation and imagination. For example, the first line, “dry vines, old trees, evening crows”, creates a sad and lonely scene. Any addition of connectives here would diminish the vividness these noun phrases offer. The poem illustrates how the Chinese language represents a high-context culture (Hall, 1976): much depends on context rather than on words. In the published English translation, however, the rules of English syntax apply. Connectives, prepositions and other functional words are needed to show the relationships between each element in the poem for clarity. These syntactic differences also uncover cultural elements: Chinese tend to employ more indirect ways of expression, whilst English favour more open and direct communication. This example demonstrates the importance of an intercultural space for mediating cultural content embedded in a language. Collections of such examples stored in parallel corpora could provide valuable learning events and promote students’ active role as intercultural mediators in translation (Granger, 2003).

Cultural deposits are further embedded in discourse, which Cook (1989: 156) defines as “stretches of language perceived to be meaningful, unified and purposive”. These stretches can be a sample of speech or writing, as short as an exchange of greetings or as long as a lecture (Song 2004). Misunderstandings in communication can occur as English and Chinese speakers often express the same idea differently. Since Chinese discourse can be highly descriptive and complex, the translator may adjust for these features by deleting information that in the target language might be seen as redundant. The following extract from a translated review of the Chinese classic The Dream of the Red Mansions is a good example (the underlined sentences in Chinese are reproduced with a line through them in English):1

1 Example taken from “Translation”: http://media.open.com.cn/media_file/200708/dongshi/fanyi/study/8/text5.htm


105

小说的艺术表现可以说是达到了出神入化的境界。严谨、缜密

的结构，生动准确的个性化语言，特别是鲜明的人物形象，都是

非常杰出的。书中出现的人物，粗计多达四百多人。不仅主角贾

宝玉、林黛玉和其他十多名主要人物成为人们熟知的艺术典型，

而且许许多多次要人物，有的甚至是一笔带过的，也都是形象鲜

明，栩栩如生。小说的艺术表现，达到如此高超的水准，在世界

文学名著中，也是极为罕见的。

It can be said that the artistic presentation of the novel has reached the acme of perfection. The novel is outstanding for its brilliantly balanced structure, a lyrical yet precise prose style and rich characterization. Although there are more than four hundred characters, the principal characters number fewer than twenty with Jia Baoyu and Lin Daiyu pre-eminent. Among the array of secondary characters, even those appearing only briefly are clearly drawn and realistic. The novel’s artistic presentation has reached a high standard that is rarely seen among literary works worldwide.

That the English translator omitted the underlined sentences in the

Chinese text in the target text reflects adaptation of the source text to the target language audience, and manifests a translation universal: simplification to reduce, for example, ambiguity and lexical density (cf. Baker 1993). In this case simplification by omission is used to adhere to the ‘KISS’ principle (Keep It Short and Simple) (Katan, 2006) which tends to be a characteristic of English expository texts. Translation, as Qian Zhongshu - a well-known Chinese translator - stated, is transformation in nature (Chen 2000), a change from one habitat to another. Cultural elements need to be included in this process of transformation because the differences often confront translators with many instances of untranslatability.

The above examples of translated texts show that when the languages (and cultures) involved differ significantly, ‘free’ translation or adaptation is required. This may pose challenges to building a parallel corpus, as the more dissimilar the languages, the harder it is to align the texts sentence by sentence, or even paragraph by paragraph. This is because omission of phrases and sentences, and modification in word order and content may have occurred. When texts are easily aligned, it suggests that the languages involved bear considerable syntactic and semantic similarity

CULTUS __________________________________________________

106

and/or, as Maia (2003) indicated, the text is skewed towards the source text. The degree of alignment between text units (word, sentence or paragraph) may thus help highlight the social and cultural elements in a source text that a translator needs to negotiate and made intelligible for the target audience.

4. Intercultural approach to translation education Clearly, translation is an intercultural activity, with “every act of

translation [being] ... an attempt at intercultural communication, a ‘negotiation space’... a privileged space for reflection” (Monti, 2009: 218). Thus the translator should be trained to function as both a bilingual mediating agent and a cultural mediator who facilitates communication, understanding, and action between people or groups from different language and cultural backgrounds.

There is an undergoing shift in teaching translation: from a linguistic approach to an intercultural approach that helps students to develop competences in both cultures. Improvement to the intercultural approach could be made by extending relevant theories/practices from language education and intercultural communication, such as the concept of ‘linguaculture’ (cf. Friedrich, 1989; and ‘languaculture’ in Agar, 1994) in the intercultural framework for teaching English in China as proposed recently by Song Li and Fu Li (2004). Linguaculture means “a complete integration of language and culture rather than language + culture because the latter cannot duly reflect the close link and interactive nature between language and culture” (2004: 30). Such integration aims at developing students’ capabilities in their first and target languages and cultures at cognitive, behavioural and interpersonal levels, and requires students “to develop a critical cultural awareness and be able to transcend linguacultural boundaries” (2004: 31).

When applied to the translation classroom, critical cultural awareness entails that trainee translators become able to evaluate practices and products of their own and the target language communities critically, without bias and with high sensitivity to linguacultural diversity. For example, when the English translation of a Chinese text adheres to the KISS principle and thus accommodates the ‘English’ reader, would a Chinese translation of an English text show similar strategies of adaptation or be closer to the norm expectation of the source text? Such


107

an approach to translation education means that teaching becomes more process-oriented rather than product-oriented: both the teacher and the learner “exploring meaning and relationships, as represented in different forms of the [source and] target language and culture” (Song and Fu 2004: 34). Exploring translation practices further raises students’ awareness of the intercultural space translators occupy and the (potential) interculturality of their conventions (cf. Pym 1993). In this process the autonomy of the learner and teacher is respected and encouraged.

Parallel corpora can be an aid to this approach by providing easily accessible collections of examples to train students on intercultural attitudes, knowledge and skills, as to be discussed in the next section. 5. Translation education: the case of parallel corpora for exploring intercultural space

The use of corpora in classrooms helps improve students’ language

competence and increase their cultural knowledge (Aston 2001). For example, KWIC (Key Word in Context) concordances of multiple corpora can reveal cultural differences in using certain words in different settings (cf. Leech and Fallon 1992), and help increase students’ knowledge of culturally connotative meanings of words and phrases in both the target and the source languages. For the translator in training, bilingual parallel corpora can be important aids. They provide “situational equivalency information” (Guidère, 2002) about the source and target text and illustrate how ideas and (inter)cultural elements are negotiated to mediate understanding. These corpora further prove their didactic values: both teachers and students explore the corpus to identify linguistic and cultural patterns in translation. Hence, the learning process can now adhere more closely to the objectives of the intercultural approach discussed above as it is no longer teacher-directed, and students are now not passive participants; instead a high degree of teacher and learner autonomy and involvement is required. Teachers are guides and facilitators for students, and create learning situations with realistic translation activities (authentic source and translated texts) in which students can develop their ‘professional self’ and become active participants in a complex communication process (cf. Kiraly 2000).

CULTUS __________________________________________________

108

Using a parallel corpus to find equivalence, however, has been more part of traditional translation practices. These focus on learning bottom-up skills to understand “source text forms and their context, and [transform] them … into linguistically ‘equivalent’ target text forms” (Vermeer 1998: 61). When words or sentences are used as translation units in a parallel corpus, the training of translators can become “restricted to the pattern of teaching fragmented sentences” (Yao 2008: 664). With the shift in paradigm from a source-oriented, small fragment approach to a more target-oriented, holistic (functional) approach, students need access to words and phrases in context (cf. Kiraly 2000). It is important to note the distinction between formal equivalence (word-for-word translation) and translational (or dynamic) equivalence. Formal equivalence “focuses attention on the message itself, in both form and content”, while dynamic equivalence is based upon “the principle of equivalent effect” (Nida 1964: 159). A static process of text transfer may increase the chance of producing a translated text based on formal equivalence. Such a text often requires additional processing from its reader. The use of a formal equivalent for the word ‘red’, with its multiple meanings and cultural connotations, may, as previously indicated, not elicit the same response or association in the target audience as the response evoked in the original audience. Hence extended meaning of a concept like ‘red’ - and its word combinations - depends on the translator’s cultural knowledge. Specifically, it requires an understanding of how to negotiate cultural space. For example, to what extent can the target text remain faithful to the source text, culture and author? At what point does faithfulness become incomprehensible to the target language audience and adaptation or freer translation is needed?

To develop this space, trainee translators need to decipher what is not explicit in a text by looking at the larger units (paragraph and even whole sections). The Parallel Corpus developed by Lixun Wang of the Hong Kong Institute of Education,2 for example, offers trainee translators the opportunity to freely access corpus files of English (nearly a million words) and Chinese (nearly half a million words) source texts. Most of the English and Chinese source texts have been compiled from classical literature and official, predominantly legal, documents that have corresponding translated versions in Chinese and English respectively. Students can thus perform concordances in English and in Chinese (in 2 http://ec-concord.ied.edu.hk/paraconc/index.htm


109

three versions: traditional, simplified and pinyin). A search for the translation of a word will further generate paragraphs in which the word occurs in the source language. This feature helps extend the meaning of the word, enabling the trainee translator to explore it in its social, cultural, historical and political contexts.

To explore the applicability of this corpus in translation education, an experiment was carried out: two students were assigned to find concordances for the two words ‘red’ and ‘wife’ using the corpus. These two words were chosen simply to showcase culturally connotative meanings. In the experiment, groups of Chinese-English translations, and English-Chinese translations were examined separately.

In the first study, 248 parallel English-Chinese translations between the English word red and its Chinese equivalent 红 were found. Among them, nearly 80% of red were literally translated into 红, indicating little connotative meanings of red in the source (English) text. For example, “Red Banners” (referring to the Chinese flag) was translated literally into 红旗, i.e. red + banners. Approximately 10% were more freely translated. The rest were either lost in translation or wrongly translated, e.g. “red-eye” was translated to “嫉妒” (meaning “jealous”). However, of the 70 parallel Chinese-English translations, only 6% of 红 were translated into its equivalent adjective red (3%) and scarlet (3%). Nearly 40% were more freely translated, e.g. “blush” for “脸红” (meaning “face turned red out of shyness”). In many cases (57%), there was no alignment with the English translations, suggesting functional adaptation rather than literal translation of the source text. The students came to the conclusion that, in comparison with 红, much of the culturally connotative meanings contained in red were omitted or went unnoticed in the Chinese translation. The students observed a lack of cultural understanding of the source language in these translations, and understood the importance of knowing both cultures in the translation process.

Translation of the word ‘wife’ is more complicated since there are a wide variety of Chinese equivalents (32 equivalents were found in the HK parallel corpus). Some (with their cultural connotations and functions) are shown in Table 1, which lists the well-defined extended meanings associated with a specific expression of “wife”. Clearly, the appropriate Chinese translation of “wife” depends on the setting (location and time) of the word use:

CULTUS __________________________________________________

110

Table 1: Selected examples of Chinese equivalents of the word “wife” and their cultural connotations and functions.

These findings prompted the students to look into the historical, social and cultural origins and meanings of the various Chinese terms for ‘wife’. Their research stimulated their intellectual curiosity, increased their knowledge of their own culture, and encouraged them to be more objective of the values and beliefs of the term as conveyed in the target language. All of these will improve the students’ capabilities to mediate, accept and explain cultural differences in future.

The above investigation exemplifies that the use of corpora encourages students to be reflective and resourceful, and helps students to develop deeper understanding of the unique and universal linguistic and cultural features of the source and target language. However, as the students worked either independently or collaboratively, the depth of analysis and learning that occurred required a degree of autonomy and curiosity on the part of the students themselves.

Zanettin (1998) further pointed out that searching for extended meaning and what is not explicit requires large corpora to reveal cultural patterns. Compared to monolingual corpora, which often exceed 300 million words, the HK parallel corpus is small but still one of the largest, if not the only, English-Chinese/Chinese-English parallel corpus containing several subtexts or genres. A recent English-Chinese addition is the Babel corpus (Xiao 2008). This corpus3 is, however, uni-directional, from English into Chinese. And it contains only single-genre texts: news articles. Its 253,633 English words and 287,462 Chinese 3 http://score.crpp.nie.edu.sg/babel/index.htm


111

tokens are significantly smaller than those in the HK corpus. But they may be more useful for investigating the degree of influence a text or genre type has on the form and lexicon of the language. For example, Yao (2008) found that in comparison to the use of bilingual dictionaries, students’ use of Babel increased their understanding of the subject matter and expanded their choice of words and expressions in translation.

6. Limitations of parallel corpora for translation education Bilingual parallel corpora can shed light on the translation choices

made and highlight the subtleties involved in making choices of translation equivalence. Currently, they still have some limitations for translation education, in particular with respect to raising awareness of the cultural and dynamic factors in translation. One major limitation is that compilers often choose texts based on technical criteria, e.g. availability and out-of-copyright. The HK parallel corpus, for example, contains mainly older material, which comes mainly from the classics. But as the intercultural approach to translation signifies, languages and cultures are in fact dynamic - they change over time at their own pace. Thus, even though a corpus of classics may well provide valuable snapshots of language use in the past, it will be less helpful for translating modern literature, texts or documents. Such corpora may thus fail to provide current cultural knowledge of the language communities.

In addition, the criteria of ‘convenience’ often limits corpus texts to governmental, legal and official sources because of their easy accessibility. Hence, this type of corpus will necessarily under represent actual language use. Another concern is that most of these parallel corpora are compiled from Indo-European language texts because of the large pools of translated material (cf. Heilbron 1999), and often provide more original texts of a globally popular language like English (as does the HK parallel corpus) than those of less international status, thus creating uni-directional corpora from English into language X. Bilingual Asian-European corpora are far less available, and in fact English-Chinese corpora, for example, are relatively sparse; and their sizes, such as that of the HK parallel corpus, are often small. This puts restrictions on the search for terms and collocations and conceals their general patterns.

CULTUS __________________________________________________

112

When working with translation corpora, this should be a concern with regard to equivalence issues. For example, Zanettin (1998) observed that translated texts fail to represent the full range of linguistic possibilities of the target language; they also reflect the stylistic idiosyncrasies of individual translators.

Uni-directional corpora like Babel cannot reveal all the cultural elements in the source language for translation effects, and limit the possibility to critically examine the degree of mediation between source and target text. With a bi-directional corpus, Wang (2008), for example, found that 75% of concessive relations in original Chinese text are expressed by implied contrastive Chinese connectors. He found also that, in the English translations 85% of these relations are expressed by explicit contrastive English connectors. Yet, when translating English into Chinese, an overt connector like however is faithfully translated into an overt Chinese equivalent seldom used in Chinese source texts. Wang (2008) concluded that an interlanguage, which is equivalent but not natural Chinese, is produced in the English-Chinese translation. This conclusion could be biased, however, since two incompatible parallel corpora were used in the study: the English-Chinese Babel corpus of recent newspaper articles and the Chinese-English corpus of the classic novel Dream of the Red Mansions. A better assessment on the degree of ‘naturalness’ in translation could certainly be made if a more recent and bi-directional corpus, compiled or balanced by genre, had been available. For example, using a comparable corpus of current original and translated Chinese texts, Xiao, He and Yue (2008) found, that in imaginative writing, press coverage, reports and official documents more conjunctions were used in the translated texts, whereas this was not necessarily the case for what Xiao et al. categorised as ‘popular lore’ (popular knowledge gained through tradition or anecdote) and academic prose. Nevertheless, Wang’s findings suggest that the intercultural space where cultures meet may not always be equally shared and will also shift over time. These important points need further investigations.

Malmkjær (1998) highlighted another problem regarding parallel corpora: only one translation for each source text. This ignores a crucial feature of translation work: being aware of and negotiating the differences between multiple translations of the same text. Consider a scenario, translator A may keep the translated work closer to the source text than translator B. Thus faithfulness constraints operate, making certain options of translation such as deletion, abbreviation, addition and


113

expansion less viable (Mansell 2008). On the other hand, translator B may operate under markedness constraints demanding that for the text to be readable certain attributes and structures in the target text need to be adhered to, and thus the translation may incorporate deletion, abbreviations and the like. The construction of corpora containing large numbers of texts and multiple translations is thus required to provide ample examples of translation strategies and to illustrate how translators have achieved equivalence under certain conditions. Such corpora not only raise awareness of general patterns, which trainee translators can observe from the recurring linguistic choices made by the translators of these texts (Yang and Li 2003), but also note the degree of interculturality revealed in these conventions.

7. Future development of parallel corpora in translation studies and education

Despite the previously discussed limitations, there have been

encouraging developments in parallel corpus activities and design. The growth of parallel corpora built and made available by higher education institutions in China and Hong Kong, and the increased interest in corpora for contrastive and translation studies in the Asian region are signs of more Asian-European corpora to materialize in the near future.

There are also other on-going efforts to increase non-English source texts in parallel corpora. For example, the structure of the English-Norwegian Parallel Corpus (ENPC), a 2.6 million word corpus, offers a good design template: providing comparable Norwegian and English source texts at the same time. The 100 original and 100 translated texts in the corpus are equally sourced from 50 comparable fiction and non-fiction texts in English and Norwegian respectively. This corpus enables (trainee) translators to observe language differences and translation effects as they can make comparisons between L1 and L2 source texts, L1 and L2 translated texts, and L1 source and L2 translated texts (cf. Johansson 1998). Having access to bidirectional corpora is important for increasing understanding of similarities and differences at the cultural level. For example, Johansson (2003) found that in fiction the English and Norwegian source texts had a major influence on the usage of love and hate and their Norwegian equivalents elske and hate respectively. In the English source texts, they were three times more common than in

CULTUS __________________________________________________

114

the Norwegian source texts, indicating that these verbs represent important cultural values in English fiction. In negotiating these values (or lack of values), in the translated texts the use of love and hate decreased in translation towards English, whereas that of elske and hate in the translation towards Norwegian increased, thereby indicating these translations closely mirrored the source texts. Johansson (2007) further found that in the source texts the Norwegian verbs collocated mainly with personal objects whereas the English verbs were most commonly paired with non-personal objects. These observations suggest that the translation effect was skewed towards the source language and that the translators were not fully aware of the difference in distribution of these verbs. Corpora like the ENPC may therefore prove themselves to be helpful tools for developing intercultural awareness in translation as students can reflect on how successful trained translators have navigated negotiable space. The design of the English-Norwegian corpus could function as a model for exploring inter-linguacultural space, and the corpus has already been replicated in other European languages, for example in English-Portuguese (COMPARA, cf. Frankenberg-Garcia and Santos 2003), but not yet in English-Chinese.

From a technical point of view, computational linguistics deserves special attention as it makes the creation of large corpora for translation a reality. Investigations in this field especially help with the building of large Chinese-English corpora, for it may provide answers to problems of, for example, part-of-speech (POS) tagging and character segmentations of Chinese texts. These problems arise from the fact that the Chinese language tends to have ambiguous POS classes with the result that words can have multiple POS tags and/or may not have one-to-one mappings with English words (Fung 1995). In addition, the lack of word delimiters, such as space and punctuation marks, to identify basic word meaning has made the segmentation of Chinese texts difficult.

Research in computer science is also producing promising results. For example, Ying Zhang, Ke Wu, Jianfeng Gao, and Phil Vine (2006) have investigated the mining of large texts and translated versions from the web. The method they employed has enabled them to identify, and automatically collect, parallel texts from the Web. Their evaluation of the obtained data set of 6500 Chinese-English candidate parallel pairs proved their method to be robust, making it promising for the future


115

development of large parallel corpora from different language systems.

8. Conclusion

Bilingual parallel corpora provide a new approach to translation

studies and training. Clearly, they can be very useful resources and a practical tool in translation, for they can be used to do contrastive linguistic research, to analyze translational transformation or translators’ style, and to raise awareness of (inter)cultural elements in translation. The use of such a corpus in the classroom can offer trainee translators a wealth of translation examples when, for example, doing searches of target words in context. In so doing, students gain access to many instances of translation for oft-used structures so as to explain and become familiar with practice. They are also able to assess how trained translators negotiate the space that can be navigated between original and translated text. These functions cannot always be fulfilled by textbooks and dictionaries.

The available parallel corpora, however, still present some obstacles because of design shortcomings and limitations in mining a wide range of bi-directional texts, and source texts with multiple translated texts. Nevertheless the work of Mona Baker and others who have studied the properties of translated texts (cf. Baker 2004), often with the help of small-sized uni-directional bilingual corpora, have illustrated their valuable use in translation education. Clearly, parallel corpora can provide interesting insights for the trainee translator into not only the linguistic differences between translated and original texts, but also the intercultural elements underlying translation.

As stated earlier, translation is a form of intercultural communication. When using parallel corpora to analyze different translation versions of the same original text, we need to incorporate an intercultural approach in the translation classroom to make students aware of the negotiable space in the translation process. Undoubtedly, the ongoing developments in parallel corpus design and mining of texts ensure bright prospects for more corpus-based research and applications in translation studies and education. Further explorations need to be made, however, on how to apply, from an intercultural perspective, a bi-directional bilingual parallel corpus such as Chinese-English effectively in the translation classroom.

CULTUS __________________________________________________

116

Acknowledgements This work is supported (through Jiang) by the Social Science

Foundation of Guangzhou City under the grant 11Y80, by the National Social Science Foundation of China under the grant 09BYY067, by the Guangdong Natural Science Foundation of China under the grant 8451009101001696, and by the International Scholar Exchange Program of Guangzhou Foreign Expert Affairs Bureau. One of us (Jiang) also thanks the US Fulbright program and Howard Community College for their support on this work. We further thank the reviewers and editor for their constructive comments and helpful suggestions.

References

Agar, M. (1994). Language shock: Understanding the culture of conversation. New York: William Morrow and Company. Aijmer, K., Altenberg, B., and Johansson, M. (eds.), 1996. Languages in

contrast: Papers from a symposium on text-based cross-linguistic studies. Lund: Lund University Press.

Aston, G. 1999. ‘Corpus use and learning to translate’. Textus, 12, 289–314.

Aston, G. (Ed.), 2001. Learning with corpora. Houston, TX: Athelstan. Baker, M. 1993. ‘Corpus linguistics and translation studies: Implications

and applications’. In M. Baker G. Francis & E. Tognini-Bonelli (Eds.) Text and technology (232-252). Philadelphia/ Amsterdam: John Benj

Baker, M. 2004. ‘A corpus-based view of similarity and difference in translation’. International Journal of Corpus Linguistics 9(2), 2004, 167-193.

Cook, G. 1989. Discourse. Oxford: Oxford University Press. Frankenberg-Garcia, A. and Santos, D. 2003. ‘Introducing Compara,

the Portuguese-English Parallel Corpus’. In F. Zanettin, S. Bernardini and D. Stewart (Eds.), Corpora in Translator Education, (pp.71-87). Manchester: St. Jerome.

Friedrich, P. 1989. ‘Language, ideology and political economy’. American Anthropologist, 91, 295-312.

Fung, P. 1995. ‘Compiling bilingual lexicon entries from a non-parallel English-Chinese corpus’. Proceedings of the Third Workshop on Very Large


117

Corpora. Retrieved 18 July 2008, from http://www.aclweb.org/anthology-new/W/W95/W95-0114.pdf

Granger, S. 2003. ‘The Corpus approach: A common way forward for Contrastive Linguistics and Translation Studies?’ In S. Granger, J. Lerot and S. Petch-Tyson (Eds.), Corpus-based approaches to contrastive linguistics and translation studies, (pp. 17-29), Amsterdam/New York: Rodopi.

Guidère, M. 2002. ‘Toward corpus-based machine translation for standard Arabic’. Translation Journal, 6 (1).

Gupta, M. 2008. Automatic acquisition of bilingual comparable corpus. Unpublished Master’s thesis, Indian Institute of Technology Kharagpur.

Hall, E. 1976. Beyond Culture. New York: Anchor Press. Heilbron, J. 1999. ‘Towards a Sociology of translation: Book translations

as a cultural world-system’. European Journal of Social Theory, 2(4), 429–444.

Jia, Y. 1997. Intercultural communication. Shanghai: Shanghai Educational Publishing House. (in Chinese)

Johansson, S. 1998. ‘On the role of corpora in cross-linguistic research’. In S. Johansson and S. Oksefjell (Eds.), Corpora and cross-linguistic research, (pp. 3-24), Amsterdam/Atlanta, GA: Rodopi.

Johansson, S. 2003. ‘Contrastive linguistics and corpora.’ In S. Granger, J. Lerot and S. Petch-Tyson (Eds.), Corpus-based approaches to contrastive linguistics and translation studies, (pp. 17-29), Amsterdam/New York: Rodopi.

Johansson, S. 2007. Seeing through multilingual corpora: On the use of corpora in contrastive studies. Amsterdam: John Benjamins B.V.

Katan, D. 2006. Translating cultures: an introduction for translators, interpreters and mediators. (《文化翻译—笔译、口译及中介入门》). Shanghai Foreign Language Education Press.

Kiraly, D. 2000. A social constructivist approach to translator education. Empowerment from theory to practice. Manchester, UK: St. Jerome Publishing.

Leech, G. and Fallon, R. (1992). ‘Computer corpora – what do they tell us about culture?’ International Computer Archive of the Modern English Journal, 16: 29-50.

Malmkjær, K. 1998. ‘Love thy neighbour: Will parallel corpora endear linguists to translators?’ Meta, 43 (4), 534–541.

CULTUS __________________________________________________

118

Maia, B. ‘The pedagogical and linguistic research implications of the GC to on-line parallel and comparable corpora’. In José João Almeida (Ed.), Corpora Paralelos, Aplicações e Algoritmos Associados, (pp. 31-32). Braga: Universidade do Minho.

Mansell, R. 2008. ‘Optimality in translation’. In A. Pym, and A. Perekrestenko (Eds.): Translation Research Projects 1, (pp. 3-12). Tarragona (Spain): Intercultural Studies Group.

Monti, E. 2009. ‘Translating the metaphors we live by’. European Journal of English Studies, 13(2), 207-221.

Nida, Eugene A. 1964). Towards a science of translating, Leiden: E. J. Brill. Puurtinen, T. 2003. ‘Genre-specific features of translationese? Linguistic

differences between translated and non-translated Finnish children’s literature’. Literary and Linguistic Computing, 18(4), 389–406.

Pym, A. 1993. ‘Why translation conventions should be intercultural rather than culture-specific. An alternative link model’. Parallèles: Cahiers de l'École de Traduction et d'Interprétation de Genève, 15, 60–68.

Schäffner, C. 2004. ‘Metaphor and translation: some implication of a cognitive approach’. Journal of Pragmatic, 36, 1253–1269.

Shuttleworth, M., and Cowie, M. 1997. Dictionary of Translation Studies. Manchester, UK: St. Jerome Publishing.

Song, L. 2004. Gateway to Intercultural Communication. Harbin: Harbin Institute of Technology Press.

Song, L and Fu, L. 2004. ‘Intercultural communicative language teaching’. EA Journal, 22(1), 20-43.

Vermeer, H.J. 1998. ‘Didactics of translation’. In M. Baker, and K. Malmkjær (Eds.): Routledge Encyclopedia of Translation Studies, (pp. 60-62). London/New York: Routledge.

Wang, J. 2008. ‘Contrastive connectors in English and Chinese: A case study of however in two translation corpora’. Corpora Proceedings of the International symposium on Using Corpora in Contrastive and Translation Studies (UCCTS2008), 25-27 September 2008, Zhejiang University, Hangzhou, China, pp. 539-560.

Xiao, R. 2008. The Babel English-Chinese parallel corpus. Retrieved October 8, 2008, from http://www.lancs.ac.uk/fass/projects/corpus/babel/babel.htm

Xiao, R., He, L., and Yue, M. 2008. ‘In pursuit of the ‘third code’: Using the ZJU Corpus of Translational Chinese in translation studies’. Proceedings of the International symposium on Using Corpora in Contrastive


119

and Translation Studies (UCCTS2008), 25-27 September 2008, Zhejiang University, Hangzhou, China, pp. 584-623.

Yang, C.C., and Li, K.W. 2003. ‘Automatic construction of English/Chinese parallel corpora’. Journal of the American Society for Information Science and technology, 54(8), 730–742.

Yao, X. 2008. ‘A corpus-based translation class’. Proceedings of the International symposium on Using Corpora in Contrastive and Translation Studies (UCCTS2008), 25-27 September 2008, Zhejiang University, Hangzhou, China, pp. 663-679.

Zanettin, F. 1998. ‘Bilingual comparable corpora and the training of translators’, In Laviosa, Sara. (Ed.): Meta, Special Issue. The corpus-based approach: a new paradigm in translation studies. 43(4), 616-630.

Zhang, Y., Wu, K., Gao, J., and Vine, P. 2006. ‘Automatic acquisition of Chinese–English parallel corpus from the Web’, in M. Lalmas et al. (Eds.): Advances in information retrieval. Proceedings of the 28th European Conference on IR Research, ECIR 200, 3936, London, UK, April 10-12, 2006, (pp. 420–431). Heidelberg: Springer-Verlag.

cultus 4 · the journal of intercultural mediation and communication corpus linguistics and culture...

Documents